Saving the whales with image processing
|
|
- Julius Hawkins
- 6 years ago
- Views:
Transcription
1 Saving the whales with image processing Youssef Ahres School of Engineering Stanford University Juhana Kangaspunta School of Engineering Stanford University I. INTRODUCTION At the moment, only 500 North Atlantic right whales are left in the world. To ensure the survival of this endangered species, marine biologists are tracking all of them to know their status and health at all times. However, manual recognition is tricky and very few researchers can perform it on the fly [1]. In this project, we implement a whale recognition system that allows researchers to reliably identify the individual whales from aerial photographs. This project is intended to be an entry to the Kaggle right whale recognition competition for which data set of labeled aerial photographs of North Atlantic right whales is provided. Throughout this paper, we will propose a whale recognition system based on two main steps: head detection and recognition. As part of this pipeline, a novel image segmentationbased filtering method is applied. We evaluate our system and compare it to the top submissions on Kaggle discussing further improvements towards the end. associated with its whale identification number. The rest of the data set is not labeled. The first challenge with this data set relates to the number of labeled images we have for a given whale. Figure 2 shows the distribution of the labeled images among the individual whales. II. DATA The Kaggle dataset consists of aerial images of North Atlantic Right Whales. Every image contains only one individual whale heading towards an arbitrary direction. Figure 1 shows an example image from the training data. Fig. 2. Histogram of whales images Fig. 1. Example image Overall, there are 447 individual whales that we would like to recognize with high confidence. To do so, we have 4545 labeled images, that is, images showing a single whale We see that most of the whales we are trying to recognize have between 1 and 10 images. This is a big limitation because the whales can have arbitrary directions and it is very hard to recognize a whale heading south based on a reference image of the same whale heading east. The head is often largely occluded hindering our capability to match keypoints or other recognition features. Finally, as we will describe in details in the next section, grayscaling these images to perform classic object detection performs very poorly. This is due to the similar grayscale intensity between the whale and the water as well as glare in the water appearing very similar to the unique identifying callosity patterns of the whale.
2 III. METHODOLOGY We approached this problem using a 2-step algorithm. First, we detect the head of the whale in the query image to reduce the dimensions and the probability of false matches due to the water and other sources of noise. Then, we use these isolated boxes to match the query image to the reference boxes. In the rest of this section, we focus on the methodology of each of these two steps. A. Head Detection Detection using cascade classifiers: To accurately detect a given object on an image, three widely used techniques seemed prominent in the literature: the Viola-Jones Haar-feature based cascade classifier[2], [13], [14] used in face detection, Histogram of Oriented Gradients (HOG) based methods [5] used in detecting humans and deep learning approaches [6], [10]. Convolutional neural networks are generally considered stateof-the-art [20]. However, they require a very large labeled data set. Lacking a simple to use implementation of a HOG detector and given the limited amount of data we had, we decided to work with the cascade classifier. In order to train this classifier, we labeled 1000 images by noting the bounding box around the whales head as positive examples and established a database of negative examples based on a random sample of water-only boxes and parts of the whales that are not intersecting the head. We also labeled the direction towards which the whale was heading. The direction is important because the cascade classifier is not rotationally invariant and is more effective when trained and optimized for a single direction using cross validation. Therefore, every query image is run through four cascade classifiers, one for each direction. This approach helped us detecting good bounding boxes for most of the images, as we will discuss in the experimental part. Filtering false positives using image segmentation: Running every query image through 4 Haar cascade classifiers increased the rate of false positives, most of which happen on the water due to various patterns of the waves that sometimes match the heads characteristics. In addition, the grayscale preprocessing step by the Haar classifier further increases the likelihood of a false positive match. Figure 3 shows an example image after grayscaling. From this image, we can see that grayscaling makes it hard to see the whale and that could easily hinder the performance of any object recognition system. Identifying the causes of false positive matches allowed us to develop an image processing pipeline to filter them out. The basic idea is to find the shape whale using image segmentation and apply it as a mask to ignore all the boxes that are not found to be on the whales body. To build such a system, we start by using K-means for image segmentation [7], [18], [19]. We attempted both RGB and HSV color spaces achieving a better separation on the HSV especially in the saturation dimension. Once the clusters are found, we can confidently assume that the largest cluster consists of the water pixels, which we set to zero, creating a noisy mask of the whale as shown below. In order to clean out the noise and have a Fig. 3. Grayscale example image Fig. 4. Mask after image segmentation using K-Means single unified whale shape in the mask, we use morphological transformation followed by object labeling. More precisely, we erode the image using a 5x5 square structured element to isolate the noise out the whale. Then, we labeled the connected components on the image and sorted them by size. The largest was considered as the whales body. This is a fundamental assumption, we are making in this approach. We will discuss its results and limitation in the experimental part of the paper. The result of this step is shown in Figure 5. Based on the Fig. 5. Mask after post processing identified whale shape, we can filter the boxes that are not overlapping enough with the mask. Formally, we define the overlap as OV ERLAP = size(b mask)/size(b) (1) Where b is the box and mask is the extracted mask.
3 We compute the defined overlap measure for every box keeping the ones that have an overlap score above This threshold was optimized empirically to maximize our filtering operation without hindering the recall. Summary: The pipeline in Figure 6 summarizes the methodology used to detect the whales head using a real example: a- Detection step using Haar features cascade classifiers b1-2) For each unlabeled image detect SURF keypoints and match the feature descriptors against all saved training set feature descriptors and compute the number of matches between the query image and the training set image. For each bounding box of the head, pick the one with the highest number of matches 3) For each unlabeled image, each individual whale is assigned a probability based on the number of matched keypoints. Fig. 7. Whale SURF features Fig. 6. Pipeline Image segmentation using K-Means in HSV color space b2erosion and connected element labeling to identify the whale c- Use the identified whale mask to filter false positive boxes Finally each unlabeled image will have a vector v of probabilities so that each element vi in the vector represents the probability of the image representing whale i. Finally to optimize the hyperparameters of the SURF detection and matching, we used cross validation on a holdout set. B. Head Recognition The goal of the project was to assign each unlabeled aerial image of a whale to a specific individual that is displayed in one or more images in the training set. Since the whales can appear in the image in various rotations, scales and with varying degrees of occlusion by the water, we needed a method that was rotationally invariant, invariant to scaling as well as robust against occlusion. Due to limited training data (approximately 10 training images per whale) we concluded that deep neural networks, especially without transfer learning, were difficult to train without overfitting. When researchers identify right whales manually, they use white callosity patterns on the whales head as the main method [11]. However, due to significant glare from the water, developing a method to extract these callosity pattern features reliably was deemed both outside the scope of this class and most likely inaccurate. Finally, we decided to use rotationally and scale invariant feature descriptors that were presented in class, namely SIFT [8], [16] and SURF [9], [17]. Initially, we experimented with SIFT, but decided to use SURF because of the faster computation time and the large number of images we would have to match against for each unlabeled image. The steps of the whale matching algorithm are as follows: 1) Detect SURF keypoints for each of the whale head bounding boxes in the training set and save the feature descriptors with the corresponding whale identifier label Fig. 8. Different whale SURF feature matches IV. E XPERIMENTAL R ESULTS In this section, we discuss the experimental setup of this algorithm and its implementation. Then, for each of the steps described above, we establish a success metric and discuss its performance. A. Head Detection This step was implemented using Matlab Image Processing toolbox and OpenCV [12] on Python. We train the Haar cascade classifiers using OpenCV (step a) and perform the image
4 B. Head Recognition The Kaggle competition submissions are evaluated using multi-class logarithmic loss which we also use to evaluate the success of our algorithm: Fig. 9. Same whale SURF feature matches segmentation (step b) using Matlab. We split the data randomly into 700 training examples and 300 testing examples. Initially, we would like to estimate the effectiveness of the filtering step. First, we determine the probability of a correctly extracted mask obtained by image segmentation and postprocessing. To do so, we compute the overlap of the ground truth bounding box and the mask following equation[1] and check whether it is above the defined threshold. If so, we consider the mask as being valid, otherwise it is not valid. Using this approach, we compute the probability of a correctly found mask and find: p = (2) Intuitively, this means that in 87.5% of the images, the approach using the K-means image segmentation and morphological transformation successfully finds the whale in the image. We also establish an evaluation metric related to the mask filtering and how many false positive boxes this approach manages to filter. To do so, we compute the number of predicted boxes on the testing set, before and after the filtering. Running this on our training set, we see that, on average, before filtering the pipeline outputs boxes per image and after filtering 9.27 boxes per image. It is reasonable since most of these boxes at least partially match the whales head. We take advantage of these partially matched boxes during the prediction step to boost our performance. Finally, to evaluate the head detection overall performance, we define the following metric: bbp = max (b boxes) size(b b gt )/size(b b gt ) (3) Where bbp stands for best box precision and b gt is the ground truth bounding box manually labeled. The intuition behind such a metric is that we use various heuristic techniques in the recognition step to take advantage of this number of boxes and aggregate them. On our testing set, we obtain: bbp = 0.78 (4) logloss = 1 N N i=1 j=1 M y ij log(p ij ) (5) In the equation N is the number of images in the test set, M is the number of whale labels, log is the natural logarithm and y ij = 1 if observation i belongs to whale j and 0 otherwise. p ij is the predicted probability that observation i belongs to whale j [1]. On an isolated test set, our algorithm achieves a logarithmic loss of which would place us in the top 30 out of 207 teams in the Kaggle competition. Unfortunately, due to time constraints and execution time challenges we were unable to run on the full Kaggle dataset and make a final submission. To put this number into perspective the Kaggle sample submission benchmark achieves a log loss of and the top submission achieves a log loss of V. CONCLUSION In this paper, we presented a methodology and an implementation for detecting and recognizing individual right whales from aerial photographs. For whale detection, the proposed image segmentation through k-means clustering in the HSV colorspace proved to be effective in filtering the high number of false positive bounding boxes arising from the Viola-Jones [2] cascade classifier. High amount of water reflection and occlusion of the whale made detection challenging but the detected bounding boxes were of sufficient precision and recall to be used in further recognition. For recognizing the right whales from the obtained bounding boxes, we proposed an approach using SURF keypoint detection and matching SURF features between the unlabeled image and training set images. Based on our initial results, this method appears to be effective in comparison to the large majority of other Kaggle competition submissions yielding a score that would earn a top 30 standing out of 207 teams. REFERENCES [1] [2] Viola P., Jones M. Rapid object detection using a boosted cascade of simple features Computer Vision and Pattern Recognition (2001) [3] Arriaga-Gomez, M.l.F., de Mendizabal-Vazquez I.; Ros-Gomez R.; Sanchez-Avila C. A comparative survey on supervised classifiers for face recognition Security Technology (ICCST), 2014 International Carnahan Conference on (2014) [4] Lu Wang; Yung, N.H.C. Extraction of Moving Objects From Their Background Based on Multiple Adaptive Thresholds and Boundary Evaluation, Intelligent Transportation Systems, IEEE Transactions on, On page(s): Volume: 11, Issue: 1, March 2010 [5] Dalal N., Triggs B. Histograms of oriented gradients for human detection Computer Vision and Pattern Recognition (2005) [6] Szegedy C., Toshev A., Dumitru E. Deep Neural Networks for Object Detection Advances in Neural Information Processing Systems 26 (NIPS 2013) [7] Tse-Wei Chen, Yi-Ling Chen, Shao-Yi Chien Fast image segmentation based on K-Means clustering with histograms in HSV color space 2008 IEEE 10th Workshop on Multimedia Signal Processing
5 [8] Lowe D.G. Object recognition from local scale-invariant features The Proceedings of the Seventh IEEE International Conference on Computer Vision, 1999 [9] Bay, Herbert, et al. Speeded-up robust features (SURF). Computer vision and image understanding (2008): [10] Girshick, Ross, et al. Rich feature hierarchies for accurate object detection and semantic [11] New England Aquarium Right Whale Callosity Pattern Identification [12] Learning OpenCV: Computer vision with the OpenCV library [13] Papageorgiou, Oren and Poggio, A general framework for object detection, International Conference on Computer Vision, 1998 [14] Lienhart, R. and Maydt, J., An extended set of Haar-like features for rapid object detection, ICIP02, pp. I: , 2002 [15] Messom, C.H. and Barczak, A.L.C., Fast and Efficient Rotated Haarlike Features Using Rotated Integral Images, Australian Conference on Robotics and Automation (ACRA2006), pp. 16, 2006 [16] Lazebnik, S., Schmid, C., and Ponce, J., Semi-Local Affine Parts for Object Recognition, BMVC, 2004 [17] P. M. Panchal, S. R. Panchal, S. K. Shah, A Comparison of SIFT and SURF, International Journal of Innovative Research in Computer and Communication Engineering Vol. 1, Issue 2, April 2013 [18] Arbelaez, Pablo, et al. Contour detection and hierarchical image segmentation. Pattern Analysis and Machine Intelligence, IEEE Transactions on 33.5 (2011): [19] Jain, Anil K. Data clustering: 50 years beyond K-means. Pattern recognition letters 31.8 (2010): [20] Krizhevsky, Alex, Ilya Sutskever, and Geoffrey E. Hinton. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems
Deep Tracking: Biologically Inspired Tracking with Deep Convolutional Networks
Deep Tracking: Biologically Inspired Tracking with Deep Convolutional Networks Si Chen The George Washington University sichen@gwmail.gwu.edu Meera Hahn Emory University mhahn7@emory.edu Mentor: Afshin
More informationReal-time Object Detection CS 229 Course Project
Real-time Object Detection CS 229 Course Project Zibo Gong 1, Tianchang He 1, and Ziyi Yang 1 1 Department of Electrical Engineering, Stanford University December 17, 2016 Abstract Objection detection
More informationImage Processing. Image Features
Image Processing Image Features Preliminaries 2 What are Image Features? Anything. What they are used for? Some statements about image fragments (patches) recognition Search for similar patches matching
More informationImage Processing Pipeline for Facial Expression Recognition under Variable Lighting
Image Processing Pipeline for Facial Expression Recognition under Variable Lighting Ralph Ma, Amr Mohamed ralphma@stanford.edu, amr1@stanford.edu Abstract Much research has been done in the field of automated
More informationCS 223B Computer Vision Problem Set 3
CS 223B Computer Vision Problem Set 3 Due: Feb. 22 nd, 2011 1 Probabilistic Recursion for Tracking In this problem you will derive a method for tracking a point of interest through a sequence of images.
More informationTRANSPARENT OBJECT DETECTION USING REGIONS WITH CONVOLUTIONAL NEURAL NETWORK
TRANSPARENT OBJECT DETECTION USING REGIONS WITH CONVOLUTIONAL NEURAL NETWORK 1 Po-Jen Lai ( 賴柏任 ), 2 Chiou-Shann Fuh ( 傅楸善 ) 1 Dept. of Electrical Engineering, National Taiwan University, Taiwan 2 Dept.
More informationSelective Search for Object Recognition
Selective Search for Object Recognition Uijlings et al. Schuyler Smith Overview Introduction Object Recognition Selective Search Similarity Metrics Results Object Recognition Kitten Goal: Problem: Where
More informationClassification of objects from Video Data (Group 30)
Classification of objects from Video Data (Group 30) Sheallika Singh 12665 Vibhuti Mahajan 12792 Aahitagni Mukherjee 12001 M Arvind 12385 1 Motivation Video surveillance has been employed for a long time
More informationSpatial Localization and Detection. Lecture 8-1
Lecture 8: Spatial Localization and Detection Lecture 8-1 Administrative - Project Proposals were due on Saturday Homework 2 due Friday 2/5 Homework 1 grades out this week Midterm will be in-class on Wednesday
More informationObject detection using non-redundant local Binary Patterns
University of Wollongong Research Online Faculty of Informatics - Papers (Archive) Faculty of Engineering and Information Sciences 2010 Object detection using non-redundant local Binary Patterns Duc Thanh
More informationDetection of a Single Hand Shape in the Foreground of Still Images
CS229 Project Final Report Detection of a Single Hand Shape in the Foreground of Still Images Toan Tran (dtoan@stanford.edu) 1. Introduction This paper is about an image detection system that can detect
More informationFace Recognition using SURF Features and SVM Classifier
International Journal of Electronics Engineering Research. ISSN 0975-6450 Volume 8, Number 1 (016) pp. 1-8 Research India Publications http://www.ripublication.com Face Recognition using SURF Features
More informationEE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm
EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm Group 1: Mina A. Makar Stanford University mamakar@stanford.edu Abstract In this report, we investigate the application of the Scale-Invariant
More informationHuman-Robot Interaction
Human-Robot Interaction Elective in Artificial Intelligence Lecture 6 Visual Perception Luca Iocchi DIAG, Sapienza University of Rome, Italy With contributions from D. D. Bloisi and A. Youssef Visual Perception
More informationAutomatic Initialization of the TLD Object Tracker: Milestone Update
Automatic Initialization of the TLD Object Tracker: Milestone Update Louis Buck May 08, 2012 1 Background TLD is a long-term, real-time tracker designed to be robust to partial and complete occlusions
More informationSelection of Scale-Invariant Parts for Object Class Recognition
Selection of Scale-Invariant Parts for Object Class Recognition Gy. Dorkó and C. Schmid INRIA Rhône-Alpes, GRAVIR-CNRS 655, av. de l Europe, 3833 Montbonnot, France fdorko,schmidg@inrialpes.fr Abstract
More informationExtracting Spatio-temporal Local Features Considering Consecutiveness of Motions
Extracting Spatio-temporal Local Features Considering Consecutiveness of Motions Akitsugu Noguchi and Keiji Yanai Department of Computer Science, The University of Electro-Communications, 1-5-1 Chofugaoka,
More informationCS231A Course Project Final Report Sign Language Recognition with Unsupervised Feature Learning
CS231A Course Project Final Report Sign Language Recognition with Unsupervised Feature Learning Justin Chen Stanford University justinkchen@stanford.edu Abstract This paper focuses on experimenting with
More informationDiscovering Visual Hierarchy through Unsupervised Learning Haider Razvi
Discovering Visual Hierarchy through Unsupervised Learning Haider Razvi hrazvi@stanford.edu 1 Introduction: We present a method for discovering visual hierarchy in a set of images. Automatically grouping
More informationLinear combinations of simple classifiers for the PASCAL challenge
Linear combinations of simple classifiers for the PASCAL challenge Nik A. Melchior and David Lee 16 721 Advanced Perception The Robotics Institute Carnegie Mellon University Email: melchior@cmu.edu, dlee1@andrew.cmu.edu
More informationAn Exploration of Computer Vision Techniques for Bird Species Classification
An Exploration of Computer Vision Techniques for Bird Species Classification Anne L. Alter, Karen M. Wang December 15, 2017 Abstract Bird classification, a fine-grained categorization task, is a complex
More informationHuman Detection and Tracking for Video Surveillance: A Cognitive Science Approach
Human Detection and Tracking for Video Surveillance: A Cognitive Science Approach Vandit Gajjar gajjar.vandit.381@ldce.ac.in Ayesha Gurnani gurnani.ayesha.52@ldce.ac.in Yash Khandhediya khandhediya.yash.364@ldce.ac.in
More informationIMAGE RETRIEVAL USING VLAD WITH MULTIPLE FEATURES
IMAGE RETRIEVAL USING VLAD WITH MULTIPLE FEATURES Pin-Syuan Huang, Jing-Yi Tsai, Yu-Fang Wang, and Chun-Yi Tsai Department of Computer Science and Information Engineering, National Taitung University,
More informationObject Detection Design challenges
Object Detection Design challenges How to efficiently search for likely objects Even simple models require searching hundreds of thousands of positions and scales Feature design and scoring How should
More informationFACE DETECTION AND RECOGNITION OF DRAWN CHARACTERS HERMAN CHAU
FACE DETECTION AND RECOGNITION OF DRAWN CHARACTERS HERMAN CHAU 1. Introduction Face detection of human beings has garnered a lot of interest and research in recent years. There are quite a few relatively
More informationResearch Article International Journals of Advanced Research in Computer Science and Software Engineering ISSN: X (Volume-7, Issue-7)
International Journals of Advanced Research in Computer Science and Software Engineering ISSN: 2277-128X (Volume-7, Issue-7) Research Article July 2017 Technique for Text Region Detection in Image Processing
More informationRotation Invariance Neural Network
Rotation Invariance Neural Network Shiyuan Li Abstract Rotation invariance and translate invariance have great values in image recognition. In this paper, we bring a new architecture in convolutional neural
More informationRecap Image Classification with Bags of Local Features
Recap Image Classification with Bags of Local Features Bag of Feature models were the state of the art for image classification for a decade BoF may still be the state of the art for instance retrieval
More informationJanitor Bot - Detecting Light Switches Jiaqi Guo, Haizi Yu December 10, 2010
1. Introduction Janitor Bot - Detecting Light Switches Jiaqi Guo, Haizi Yu December 10, 2010 The demand for janitorial robots has gone up with the rising affluence and increasingly busy lifestyles of people
More informationA Novel Extreme Point Selection Algorithm in SIFT
A Novel Extreme Point Selection Algorithm in SIFT Ding Zuchun School of Electronic and Communication, South China University of Technolog Guangzhou, China zucding@gmail.com Abstract. This paper proposes
More informationDeformable Part Models
CS 1674: Intro to Computer Vision Deformable Part Models Prof. Adriana Kovashka University of Pittsburgh November 9, 2016 Today: Object category detection Window-based approaches: Last time: Viola-Jones
More informationComputer and Machine Vision
Computer and Machine Vision Lecture Week 10 Part-2 Skeletal Models and Face Detection March 21, 2014 Sam Siewert Outline of Week 10 Lab #4 Overview Lab #5 and #6 Extended Lab Overview SIFT and SURF High
More informationCS 231A Computer Vision (Fall 2012) Problem Set 3
CS 231A Computer Vision (Fall 2012) Problem Set 3 Due: Nov. 13 th, 2012 (2:15pm) 1 Probabilistic Recursion for Tracking (20 points) In this problem you will derive a method for tracking a point of interest
More informationMouse Pointer Tracking with Eyes
Mouse Pointer Tracking with Eyes H. Mhamdi, N. Hamrouni, A. Temimi, and M. Bouhlel Abstract In this article, we expose our research work in Human-machine Interaction. The research consists in manipulating
More informationTHE SPEED-LIMIT SIGN DETECTION AND RECOGNITION SYSTEM
THE SPEED-LIMIT SIGN DETECTION AND RECOGNITION SYSTEM Kuo-Hsin Tu ( 塗國星 ), Chiou-Shann Fuh ( 傅楸善 ) Dept. of Computer Science and Information Engineering, National Taiwan University, Taiwan E-mail: p04922004@csie.ntu.edu.tw,
More informationHuman Detection. A state-of-the-art survey. Mohammad Dorgham. University of Hamburg
Human Detection A state-of-the-art survey Mohammad Dorgham University of Hamburg Presentation outline Motivation Applications Overview of approaches (categorized) Approaches details References Motivation
More informationWindow based detectors
Window based detectors CS 554 Computer Vision Pinar Duygulu Bilkent University (Source: James Hays, Brown) Today Window-based generic object detection basic pipeline boosting classifiers face detection
More informationRecognizing Apples by Piecing Together the Segmentation Puzzle
Recognizing Apples by Piecing Together the Segmentation Puzzle Kyle Wilshusen 1 and Stephen Nuske 2 Abstract This paper presents a system that can provide yield estimates in apple orchards. This is done
More informationHand Posture Recognition Using Adaboost with SIFT for Human Robot Interaction
Hand Posture Recognition Using Adaboost with SIFT for Human Robot Interaction Chieh-Chih Wang and Ko-Chih Wang Department of Computer Science and Information Engineering Graduate Institute of Networking
More informationKey properties of local features
Key properties of local features Locality, robust against occlusions Must be highly distinctive, a good feature should allow for correct object identification with low probability of mismatch Easy to etract
More informationObject detection using Region Proposals (RCNN) Ernest Cheung COMP Presentation
Object detection using Region Proposals (RCNN) Ernest Cheung COMP790-125 Presentation 1 2 Problem to solve Object detection Input: Image Output: Bounding box of the object 3 Object detection using CNN
More informationLarge-Scale Traffic Sign Recognition based on Local Features and Color Segmentation
Large-Scale Traffic Sign Recognition based on Local Features and Color Segmentation M. Blauth, E. Kraft, F. Hirschenberger, M. Böhm Fraunhofer Institute for Industrial Mathematics, Fraunhofer-Platz 1,
More informationShort Survey on Static Hand Gesture Recognition
Short Survey on Static Hand Gesture Recognition Huu-Hung Huynh University of Science and Technology The University of Danang, Vietnam Duc-Hoang Vo University of Science and Technology The University of
More informationDeterminant of homography-matrix-based multiple-object recognition
Determinant of homography-matrix-based multiple-object recognition 1 Nagachetan Bangalore, Madhu Kiran, Anil Suryaprakash Visio Ingenii Limited F2-F3 Maxet House Liverpool Road Luton, LU1 1RS United Kingdom
More informationhttps://en.wikipedia.org/wiki/the_dress Recap: Viola-Jones sliding window detector Fast detection through two mechanisms Quickly eliminate unlikely windows Use features that are fast to compute Viola
More informationDistance-Based Descriptors and Their Application in the Task of Object Detection
Distance-Based Descriptors and Their Application in the Task of Object Detection Radovan Fusek (B) and Eduard Sojka Department of Computer Science, Technical University of Ostrava, FEECS, 17. Listopadu
More informationImage-Based Competitive Printed Circuit Board Analysis
Image-Based Competitive Printed Circuit Board Analysis Simon Basilico Department of Electrical Engineering Stanford University Stanford, CA basilico@stanford.edu Ford Rylander Department of Electrical
More informationDeep Learning in Visual Recognition. Thanks Da Zhang for the slides
Deep Learning in Visual Recognition Thanks Da Zhang for the slides Deep Learning is Everywhere 2 Roadmap Introduction Convolutional Neural Network Application Image Classification Object Detection Object
More informationPreviously. Window-based models for generic object detection 4/11/2011
Previously for generic object detection Monday, April 11 UT-Austin Instance recognition Local features: detection and description Local feature matching, scalable indexing Spatial verification Intro to
More informationFeature descriptors. Alain Pagani Prof. Didier Stricker. Computer Vision: Object and People Tracking
Feature descriptors Alain Pagani Prof. Didier Stricker Computer Vision: Object and People Tracking 1 Overview Previous lectures: Feature extraction Today: Gradiant/edge Points (Kanade-Tomasi + Harris)
More informationTraffic Signs Recognition using HP and HOG Descriptors Combined to MLP and SVM Classifiers
Traffic Signs Recognition using HP and HOG Descriptors Combined to MLP and SVM Classifiers A. Salhi, B. Minaoui, M. Fakir, H. Chakib, H. Grimech Faculty of science and Technology Sultan Moulay Slimane
More informationDiscriminative classifiers for image recognition
Discriminative classifiers for image recognition May 26 th, 2015 Yong Jae Lee UC Davis Outline Last time: window-based generic object detection basic pipeline face detection with boosting as case study
More informationMULTI ORIENTATION PERFORMANCE OF FEATURE EXTRACTION FOR HUMAN HEAD RECOGNITION
MULTI ORIENTATION PERFORMANCE OF FEATURE EXTRACTION FOR HUMAN HEAD RECOGNITION Panca Mudjirahardjo, Rahmadwati, Nanang Sulistiyanto and R. Arief Setyawan Department of Electrical Engineering, Faculty of
More informationConvolutional Neural Networks. Computer Vision Jia-Bin Huang, Virginia Tech
Convolutional Neural Networks Computer Vision Jia-Bin Huang, Virginia Tech Today s class Overview Convolutional Neural Network (CNN) Training CNN Understanding and Visualizing CNN Image Categorization:
More informationReal-Time Model-Based Hand Localization for Unsupervised Palmar Image Acquisition
Real-Time Model-Based Hand Localization for Unsupervised Palmar Image Acquisition Ivan Fratric 1, Slobodan Ribaric 1 1 University of Zagreb, Faculty of Electrical Engineering and Computing, Unska 3, 10000
More informationEfficient Segmentation-Aided Text Detection For Intelligent Robots
Efficient Segmentation-Aided Text Detection For Intelligent Robots Junting Zhang, Yuewei Na, Siyang Li, C.-C. Jay Kuo University of Southern California Outline Problem Definition and Motivation Related
More informationPedestrian Detection and Tracking in Images and Videos
Pedestrian Detection and Tracking in Images and Videos Azar Fazel Stanford University azarf@stanford.edu Viet Vo Stanford University vtvo@stanford.edu Abstract The increase in population density and accessibility
More informationCS4758: Moving Person Avoider
CS4758: Moving Person Avoider Yi Heng Lee, Sze Kiat Sim Abstract We attempt to have a quadrotor autonomously avoid people while moving through an indoor environment. Our algorithm for detecting people
More informationImageNet Classification with Deep Convolutional Neural Networks
ImageNet Classification with Deep Convolutional Neural Networks Alex Krizhevsky Ilya Sutskever Geoffrey Hinton University of Toronto Canada Paper with same name to appear in NIPS 2012 Main idea Architecture
More informationProject Report for EE7700
Project Report for EE7700 Name: Jing Chen, Shaoming Chen Student ID: 89-507-3494, 89-295-9668 Face Tracking 1. Objective of the study Given a video, this semester project aims at implementing algorithms
More informationScott Smith Advanced Image Processing March 15, Speeded-Up Robust Features SURF
Scott Smith Advanced Image Processing March 15, 2011 Speeded-Up Robust Features SURF Overview Why SURF? How SURF works Feature detection Scale Space Rotational invariance Feature vectors SURF vs Sift Assumptions
More informationMobile Face Recognization
Mobile Face Recognization CS4670 Final Project Cooper Bills and Jason Yosinski {csb88,jy495}@cornell.edu December 12, 2010 Abstract We created a mobile based system for detecting faces within a picture
More informationSURF. Lecture6: SURF and HOG. Integral Image. Feature Evaluation with Integral Image
SURF CSED441:Introduction to Computer Vision (2015S) Lecture6: SURF and HOG Bohyung Han CSE, POSTECH bhhan@postech.ac.kr Speed Up Robust Features (SURF) Simplified version of SIFT Faster computation but
More informationEpithelial rosette detection in microscopic images
Epithelial rosette detection in microscopic images Kun Liu,3, Sandra Ernst 2,3, Virginie Lecaudey 2,3 and Olaf Ronneberger,3 Department of Computer Science 2 Department of Developmental Biology 3 BIOSS
More informationCS 231A Computer Vision (Winter 2014) Problem Set 3
CS 231A Computer Vision (Winter 2014) Problem Set 3 Due: Feb. 18 th, 2015 (11:59pm) 1 Single Object Recognition Via SIFT (45 points) In his 2004 SIFT paper, David Lowe demonstrates impressive object recognition
More informationGeneric Face Alignment Using an Improved Active Shape Model
Generic Face Alignment Using an Improved Active Shape Model Liting Wang, Xiaoqing Ding, Chi Fang Electronic Engineering Department, Tsinghua University, Beijing, China {wanglt, dxq, fangchi} @ocrserv.ee.tsinghua.edu.cn
More informationChapter 3 Image Registration. Chapter 3 Image Registration
Chapter 3 Image Registration Distributed Algorithms for Introduction (1) Definition: Image Registration Input: 2 images of the same scene but taken from different perspectives Goal: Identify transformation
More informationDesigning Applications that See Lecture 7: Object Recognition
stanford hci group / cs377s Designing Applications that See Lecture 7: Object Recognition Dan Maynes-Aminzade 29 January 2008 Designing Applications that See http://cs377s.stanford.edu Reminders Pick up
More informationA FRAMEWORK OF EXTRACTING MULTI-SCALE FEATURES USING MULTIPLE CONVOLUTIONAL NEURAL NETWORKS. Kuan-Chuan Peng and Tsuhan Chen
A FRAMEWORK OF EXTRACTING MULTI-SCALE FEATURES USING MULTIPLE CONVOLUTIONAL NEURAL NETWORKS Kuan-Chuan Peng and Tsuhan Chen School of Electrical and Computer Engineering, Cornell University, Ithaca, NY
More informationUnderstanding Tracking and StroMotion of Soccer Ball
Understanding Tracking and StroMotion of Soccer Ball Nhat H. Nguyen Master Student 205 Witherspoon Hall Charlotte, NC 28223 704 656 2021 rich.uncc@gmail.com ABSTRACT Soccer requires rapid ball movements.
More informationRushes Video Segmentation Using Semantic Features
Rushes Video Segmentation Using Semantic Features Athina Pappa, Vasileios Chasanis, and Antonis Ioannidis Department of Computer Science and Engineering, University of Ioannina, GR 45110, Ioannina, Greece
More informationToward Part-based Document Image Decoding
2012 10th IAPR International Workshop on Document Analysis Systems Toward Part-based Document Image Decoding Wang Song, Seiichi Uchida Kyushu University, Fukuoka, Japan wangsong@human.ait.kyushu-u.ac.jp,
More informationCS 231A Computer Vision (Fall 2011) Problem Set 4
CS 231A Computer Vision (Fall 2011) Problem Set 4 Due: Nov. 30 th, 2011 (9:30am) 1 Part-based models for Object Recognition (50 points) One approach to object recognition is to use a deformable part-based
More informationTraffic Sign Localization and Classification Methods: An Overview
Traffic Sign Localization and Classification Methods: An Overview Ivan Filković University of Zagreb Faculty of Electrical Engineering and Computing Department of Electronics, Microelectronics, Computer
More informationINTELLIGENT transportation systems have a significant
INTL JOURNAL OF ELECTRONICS AND TELECOMMUNICATIONS, 205, VOL. 6, NO. 4, PP. 35 356 Manuscript received October 4, 205; revised November, 205. DOI: 0.55/eletel-205-0046 Efficient Two-Step Approach for Automatic
More informationVideo Processing for Judicial Applications
Video Processing for Judicial Applications Konstantinos Avgerinakis, Alexia Briassouli, Ioannis Kompatsiaris Informatics and Telematics Institute, Centre for Research and Technology, Hellas Thessaloniki,
More information11. Image Data Analytics. Jacobs University Visualization and Computer Graphics Lab
11. Image Data Analytics Motivation Images (and even videos) have become a popular data format for storing information digitally. Data Analytics 377 Motivation Traditionally, scientific and medical imaging
More informationTri-modal Human Body Segmentation
Tri-modal Human Body Segmentation Master of Science Thesis Cristina Palmero Cantariño Advisor: Sergio Escalera Guerrero February 6, 2014 Outline 1 Introduction 2 Tri-modal dataset 3 Proposed baseline 4
More informationImageCLEF 2011
SZTAKI @ ImageCLEF 2011 Bálint Daróczy joint work with András Benczúr, Róbert Pethes Data Mining and Web Search Group Computer and Automation Research Institute Hungarian Academy of Sciences Training/test
More informationThe Detection of Faces in Color Images: EE368 Project Report
The Detection of Faces in Color Images: EE368 Project Report Angela Chau, Ezinne Oji, Jeff Walters Dept. of Electrical Engineering Stanford University Stanford, CA 9435 angichau,ezinne,jwalt@stanford.edu
More informationRETRIEVAL OF FACES BASED ON SIMILARITIES Jonnadula Narasimha Rao, Keerthi Krishna Sai Viswanadh, Namani Sandeep, Allanki Upasana
ISSN 2320-9194 1 Volume 5, Issue 4, April 2017, Online: ISSN 2320-9194 RETRIEVAL OF FACES BASED ON SIMILARITIES Jonnadula Narasimha Rao, Keerthi Krishna Sai Viswanadh, Namani Sandeep, Allanki Upasana ABSTRACT
More informationToward Retail Product Recognition on Grocery Shelves
Toward Retail Product Recognition on Grocery Shelves Gül Varol gul.varol@boun.edu.tr Boğaziçi University, İstanbul, Turkey İdea Teknoloji Çözümleri, İstanbul, Turkey Rıdvan S. Kuzu ridvan.salih@boun.edu.tr
More informationFacial Feature Extraction Based On FPD and GLCM Algorithms
Facial Feature Extraction Based On FPD and GLCM Algorithms Dr. S. Vijayarani 1, S. Priyatharsini 2 Assistant Professor, Department of Computer Science, School of Computer Science and Engineering, Bharathiar
More informationInternational Journal of Advanced Research in Computer Science and Software Engineering
ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: Fingerprint Recognition using Robust Local Features Madhuri and
More informationSUMMARY: DISTINCTIVE IMAGE FEATURES FROM SCALE- INVARIANT KEYPOINTS
SUMMARY: DISTINCTIVE IMAGE FEATURES FROM SCALE- INVARIANT KEYPOINTS Cognitive Robotics Original: David G. Lowe, 004 Summary: Coen van Leeuwen, s1460919 Abstract: This article presents a method to extract
More informationA Robust Hand Gesture Recognition Using Combined Moment Invariants in Hand Shape
, pp.89-94 http://dx.doi.org/10.14257/astl.2016.122.17 A Robust Hand Gesture Recognition Using Combined Moment Invariants in Hand Shape Seungmin Leem 1, Hyeonseok Jeong 1, Yonghwan Lee 2, Sungyoung Kim
More informationBus Detection and recognition for visually impaired people
Bus Detection and recognition for visually impaired people Hangrong Pan, Chucai Yi, and Yingli Tian The City College of New York The Graduate Center The City University of New York MAP4VIP Outline Motivation
More informationObject Detection Using Segmented Images
Object Detection Using Segmented Images Naran Bayanbat Stanford University Palo Alto, CA naranb@stanford.edu Jason Chen Stanford University Palo Alto, CA jasonch@stanford.edu Abstract Object detection
More informationFace Detection using Hierarchical SVM
Face Detection using Hierarchical SVM ECE 795 Pattern Recognition Christos Kyrkou Fall Semester 2010 1. Introduction Face detection in video is the process of detecting and classifying small images extracted
More informationAutomatic License Plate Detection
Automatic License Plate Detection CS771 Course Project Winter Semester 2015-16 Author: Anurag Sharma(12146) Anurendra Kumar(12147) K.V Sameer Raja(12332) Shreesh Ladha(12679) Supervisors: Prof Harish Karnick
More informationWeighted Convolutional Neural Network. Ensemble.
Weighted Convolutional Neural Network Ensemble Xavier Frazão and Luís A. Alexandre Dept. of Informatics, Univ. Beira Interior and Instituto de Telecomunicações Covilhã, Portugal xavierfrazao@gmail.com
More informationLOCAL AND GLOBAL DESCRIPTORS FOR PLACE RECOGNITION IN ROBOTICS
8th International DAAAM Baltic Conference "INDUSTRIAL ENGINEERING - 19-21 April 2012, Tallinn, Estonia LOCAL AND GLOBAL DESCRIPTORS FOR PLACE RECOGNITION IN ROBOTICS Shvarts, D. & Tamre, M. Abstract: The
More informationFace and Nose Detection in Digital Images using Local Binary Patterns
Face and Nose Detection in Digital Images using Local Binary Patterns Stanko Kružić Post-graduate student University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
More informationThe most cited papers in Computer Vision
COMPUTER VISION, PUBLICATION The most cited papers in Computer Vision In Computer Vision, Paper Talk on February 10, 2012 at 11:10 pm by gooly (Li Yang Ku) Although it s not always the case that a paper
More informationExploiting scene constraints to improve object detection algorithms for industrial applications
Exploiting scene constraints to improve object detection algorithms for industrial applications PhD Public Defense Steven Puttemans Promotor: Toon Goedemé 2 A general introduction Object detection? Help
More informationA New Strategy of Pedestrian Detection Based on Pseudo- Wavelet Transform and SVM
A New Strategy of Pedestrian Detection Based on Pseudo- Wavelet Transform and SVM M.Ranjbarikoohi, M.Menhaj and M.Sarikhani Abstract: Pedestrian detection has great importance in automotive vision systems
More information2D Image Processing Feature Descriptors
2D Image Processing Feature Descriptors Prof. Didier Stricker Kaiserlautern University http://ags.cs.uni-kl.de/ DFKI Deutsches Forschungszentrum für Künstliche Intelligenz http://av.dfki.de 1 Overview
More informationFace Detection on OpenCV using Raspberry Pi
Face Detection on OpenCV using Raspberry Pi Narayan V. Naik Aadhrasa Venunadan Kumara K R Department of ECE Department of ECE Department of ECE GSIT, Karwar, Karnataka GSIT, Karwar, Karnataka GSIT, Karwar,
More informationEdge Detection Using Convolutional Neural Network
Edge Detection Using Convolutional Neural Network Ruohui Wang (B) Department of Information Engineering, The Chinese University of Hong Kong, Hong Kong, China wr013@ie.cuhk.edu.hk Abstract. In this work,
More informationRobust PDF Table Locator
Robust PDF Table Locator December 17, 2016 1 Introduction Data scientists rely on an abundance of tabular data stored in easy-to-machine-read formats like.csv files. Unfortunately, most government records
More information